New error swap_full node 1

Support
, ,
1

NethServer Version: ns8
Module: Core?

In the past 24 hours I started to get swap_full errors. I’m trying to figure out where this is coming from and the cause but not having much luck so far.

Here is the error email:

image
image1104×798 60.4 KB

I’ve found this is the Node Details tab of the Cluster login

image
image2274×970 130 KB

So I’m assuming that my NS8 (which is running on Proxmox) is running out of swap.

I can’t find anything in the docs that talks to swap size or requirements.

Is there guidance somewhere on what should be set for swap in the NS8 core and how to set it?

Thanks in advance.

2

In short, 1GB is too small, you’d allocate 4GB at least.

The best reference to start is the base OS official documentation. This is for Rocky Linux, but there is one for Debian for sure: Chapter 15. Getting started with swap | Managing storage devices | Red Hat Enterprise Linux | 9 | Red Hat Documentation

Edit:

For swap file on SSDs, you may want to tune swappiness. Please, refer to this article: How To Add Swap Space on Ubuntu 22.04 | DigitalOcean

:warning: fallocate may be dangerous to use in some environments, because it generates a faulty swap file. The dd command is generally slower but safer.

2 Likes
3

IMVHO memory is also too small - I’d say, at least 32 GB even better >64 GB… little depending what you are orchestrating with your IT.

4

Thank you @davidep and @schulzstefan .

I created the original VM exactly the same as what I had for NS7. The VM runs on an SSD so thought it should be OK. I didn’t think it needed any additional tinkering.

I will read the links that you shared so I can understand it better and will also look at @Andy_Wismer’s post about expanding the VM disk and adding swap as his instructions were for Debian.

I’ll need to do some searching on the swappiness for SSDs.

Thanks for your help and suggestions.

2 Likes